Research On Infrared And Visible Image Preprocessing Technology

This topic comes from the sub-topic 5 "Research and Development of Shipboard Infrared and Visible Sea Distress Target Searching System" of Project Four of the Ministry of Industry and Information Technology "Research and Development of Shipborne Sea Distress Target Quick Search Technology and Device Development".In view of the complex marine environment in which large and deep-sea rescue ships are located,using infrared and visible light detection imaging equipment,etc.,by building a rescue ship-borne sea distress target search system,it can quickly and widely search for sea distress targets,improve the detection of sea distress targets,Positioning and tracking capabilities,and thus improve the success rate of search and rescue of China's long-distance rescue ship under the complex sea conditions.Image pre-processing technology optimizes the quality degradation of infrared and visible light images,improves image quality,and provides technical support for subsequent research on image registration and target detection.The research of image preprocessing technology in this paper mainly includes two parts:infrared image contrast enhancement and infrared image deblurring algorithm research,and visible channel image removal of color distortion dark channel defogging algorithm research.Infrared image contrast enhancement algorithm research mainly focuses on the problems of low contrast,unclear texture details and uneven brightness in the original infrared image.A contrast enhancement algorithm that combines edge information is proposed.This method combines layered processing and enhanced filtering.Combined to achieve contrast enhancement of infrared images.The first step is layered processing.The guided infrared filter is used to layer the original infrared image to obtain the basic sub-image and detail sub-image ISD.The two sub-images after processing are fused into a pioture;the second step of enhancement filtering is to use adaptive histogram.equalization and Laplacian sharpening filtering to process the original infrared image and generate a picture ILaplace;finally,the previous The two-step generated imagesILaplace and images ISD are weighted and fused to obtain the final contrast-enhanced image.This image can overcome' the "over-enhancement"and gradient inversion phenomena,and the contrast and texture details have been significantly improved.The research of infrared image deblurring algorithm is mainly aimed at the problem of blurring details in the original infrared image.On the basis of the classic deblurring algorithm Lucy-Richard(RL)algorithm,an improved deblurring based on RL is proposed The algorithm introduces local priors and auxiliary graphs in the iterative process of the RL algorithm,uses local priors to suppress the ringing effect in smooth areas of the image,and uses auxiliary graphs to prevent noise amplification during the iteration process;by selecting reasonable parameters The algorithm has achieved good results in suppressing the ringing effect propagation and noise amplification and ensuring the completeness of the image details.Visible light defogging algorithm research is mainly aimed at the problem of color distortion in the image after He Kai's bright and dark channel prior defogging algorithm processing.Introducing a boundary parameter in the He Kai light-dark channel prior defogging formula,so that the improved defogging algorithm can more accurately calculate the transmittance of bright white areas that do not satisfy the dark channel prior assumption,avoiding fogging visible light images from passing through the fog The color distortion produced after the treatment significantly improves the overall visual effect of the processed visible light image.The effectiveness of the improved defogging algorithm in removing color distortion is verified through experiments.In order to verify the effectiveness of the three image preprocessing algorithms studied in this paper,the appropriate infrared images and visible light images were selected to conduct comparative experiments on the three image preprocessing algorithms,respectively.The subjective visual effects evaluation and objective quality evaluation of the algorithm The effectiveness was verified and evaluated.It is confirmed by experiments that the three image preprocessing algorithms studied in this paper can achieve very good processing effects,and have obvious advantages compared with other image preprocessing algorithms.